Specific identification of butanol isomers via MOF-functionalized on-chip photonic waveguide sensor arrays

Accurate identification of volatile organic compounds (VOCs) in gas mixtures remains a significant challenge, especially for isomers with similar structures. We propose a metal-organic framework (MOF)-functionalized micro-ring resonator (MRR) array for butanol isomer recognition and mixture analysis. By growing HKUST-1 thin films on MRRs with diverse waveguide geometries using spin-assisted liquid-phase epitaxial layer-by-layer (SA-LPE-LbL) methods, the optimized racetrack-type MRR achieves 17.7 pm/ppm sensitivity and a calculated 0.41 ppm detection limit for 1-butanol. Additionally, the MRR array demonstrated reliable environmental tolerance under near-realistic humidity conditions (30–50% RH). While sensitivity slightly decreased, the ability to distinguish isomers remains effective. Molecular dynamics (MD) simulations reveal that steric hindrance governs the adsorption differences. Specifically, linear isomer access Cu^{2 +} sites more effectively in HKUST-1, leading to stronger binding and faster diffusion. Deep learning models analyze the multi-channel spectral data to achieve a 94.4% classification accuracy for the three isomers and effectively decompose gas mixture spectra with a per-wavelength-point (PWP) reconstruction error below 0.04. This platform significantly enhances sensitivity and selectivity for isomer differentiation and mixture analysis, offering a scalable solution for optical gas sensing.